Intake of Legumes and the Risk of Cardiovascular Disease: Frailty Modeling of a Prospective Cohort Study in the Iranian Middle-Aged and Older Population

Home , Counties of Iran

ORIGINAL ARTICLE Intake of legumes and the risk of cardiovascular disease: frailty modeling of a prospective cohort study in the Iranian middle-aged and older population

F Nouri1,2, N Sarrafzadegan1, N Mohammadifard3, M Sadeghi4 and M Mansourian2

BACKGROUND/OBJECTIVES: The purpose of this study was to explore the association of legume intake (beans, chickpeas, lentils and so on), as part of a low-glycemic index diet, with the risk of cardiovascular events in the Iranian middle- and old-aged people. SUBJECTS/METHODS: A total of 6504 subjects living in the three counties of Iran participated in the Isfahan Cohort Study. Totally, 6323 were free of cardiovascular disease (CVD) at their baseline examination. Of the 6323 individuals, 5398 participants remained in the study for 7 years of follow-up. They have been contacted every 2 years for possible occurrence of CVD events including fatal and non-fatal myocardial infarction, unstable angina, fatal and non-fatal stroke, and sudden cardiac death. The frequency of legume intake was estimated using a food frequency questionnaire. Cox proportional hazards models with shared gamma frailty terms were used to model time to event outcomes. RESULTS: After a median follow-up of 6.8 years, 427 cardiovascular events occurred. The intake of legumes in different tertiles of consuming measure was associated with 34% lower risk of CVD in old-aged people, after controlling for the other probable confounders (hazard ratio and 95% CI: 0.66 (0.45, 0.98), P-value = 0.039). However, there was no signiﬁcant association between the frequency of consuming legumes and CVD events in the middle-aged people. CONCLUSIONS: The present study indicated a strong inverse relationship between legume intake and the risk of cardiovascular events in old-aged Iranian people. European Journal of Clinical Nutrition (2016) 70, 217–221; doi:10.1038/ejcn.2015.153; published online 23 September 2015

INTRODUCTION decreased oxidized low-density lipoprotein levels. These studies Cardiovascular disease (CVD) is the leading cause of death performed in people of different age groups indicated that the risk worldwide.1 CVD is responsible for 430% of global deaths.2 of CVD reduced in people who consumed a diet containing at 22 According to a recent World Health Organization report, at least least 37 g of dietary fiber per day. It is thus necessary to 80% of CVD morbidities and mortalities could be prevented by investigate the effects of diets, especially, in older populations. improving modifiable CVD risk factors.3 The role of diet, an The important role of diet in the prevention of disability and important component of lifestyle, is crucial in the development premature death in the older populations is a well-recognized and and prevention of CVD.4 Although nutritional recommendations accepted means of geriatric care. In addition, due to the limited available studies investigating diet and CVD risk in elders, findings are easier to follow when phrased in terms of the consumption of 23 foods rather than nutrients,5 most studies have focused on the from middle-aged populations steer older adult guidelines. To our knowledge, there has not been any report regarding the associations of CVD risk with nutrients. Epidemiological data have fl clearly shown a negative relationship between plant-based food in uence of legume intake on risk of cardiovascular events in consumption, such as legumes, and CVD risk.6,7 Legumes Iranian middle-aged and older population. Thus, the purpose of including beans, chickpeas and lentils are potentially valuable this study was to explore the association of legume intake with the risk of CVD occurrence in middle-aged and older people who dietary sources of low sodium and rich potassium, calcium, participated in Isfahan Cohort Study (ICS).24 magnesium and fiber.6 Numerous studies have indicated that the nutrients in beans, for example, fiber,7,8 folate,9 magnesium10,11 and copper,12 have inverse associations with CVD.13–17 Some MATERIALS AND METHODS studies have shown that whole grain and legumes, containing 24 high contents of fiber, are known to be slowly digested and The ICS is a community-based longitudinal ongoing study of adults aged ⩾ absorbed; therefore, they have relatively low glycemic indexes 35 years that started in 2001. Participants living in urban and rural areas 18,19 of three counties in central Iran (Isfahan, Najaf Abad and Arak) who had associated with lower concentration of insulin. Also, legumes participated in the baseline study of Isfahan Healthy Heart Program (IHHP) – 25 are good sources of protein (20 30%) compared with usual plant were enrolled in ICS. IHHP was a community trial for CVD prevention and protein sources (o20%) and high protein in legumes improves control conducted by Isfahan Cardiovascular Research Center (ICRC), a antioxidant response,20,21 which has been associated with World Health Organization (WHO) collaborating center in the eastern

1Isfahan Cardiovascular Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran; 2Department of Epidemiology and Biostatistics, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran; 3Hypertension Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran and 4Cardiac Rehabilitation Research Center, Cardiovascular Research Institute, Isfahan University of Medical Sciences, Isfahan, Iran. Correspondence: Dr M Mansourian, Department of Epidemiology and Biostatistics, School of Health, Isfahan University of Medical Sciences, Isfahan, Iran. E-mail: [email protected] Received 18 June 2014; revised 22 July 2015; accepted 3 August 2015; published online 23 September 2015 Association of legume intake with cardiovascular disease risk F Nouri et al 218 Mediterranean region. The IHHP sample was recruited in different age and independent variables and clusters of urban or rural areas of the three sex groups. The prevalence of cardiovascular risk factors was estimated to districts were taken into account as the shared frailty part. The hazard be 0.2 and the clustering sample size was determined by the power of 90% ratios and 95% confidence intervals (CI) for CVD event are presented in with a relative risk of 0.75 at a significance of 0.05. The overall sample size three different model strategies (unadjusted model; adjusted model by was 4828 in each area, allowing for losses to follow-up it was intended to sex, age, income, marital status, education, job, ever smoke, leisure-time recruit 6300 in each area. A total of 12 514 individuals were included in the physical activity, legume-adjusted dietary score (using residual model), baseline survey of IHHP. Among 12 514 individuals, there were 6504 adults anti-hyperlipidemia medicine; and further adjusted for BMI). The propor- aged ⩾ 35 years who participated in the ICS. The participants were tional hazards assumption was checked using suitable graphical proce- recruited from 2 January to 28 September 2001 using multistage random dures or goodness of fit test. P-values o0.05 (two-tailed) were considered cluster sampling. Details of the IHHP design have been previously as statistically significant. reported.24–26 Ethical approval was obtained from the Ethics Committee of ICRC. After obtaining informed written consent, participants completed standardized questionnaires on health status, medical history and RESULTS cardiovascular risk factors, and underwent clinical examination, electro- cardiography and laboratory evaluation. Whole information about the After a median follow-up of 6.8 years, 5398 participants remained individuals was collected by trained interviewers. Participants with known in the study and 427 CVD events occurred. It consisted of 89 coronary heart disease, heart failure, stroke and ischemic heart attack were (20.8%) cases of myocardial infarction, 91 (21.3%) cases of stroke, excluded from the study (n = 181). Among baseline participants, 5398 54 (12.6%) cases of sudden cardiac death and 193 (45.1%) cases of participants remained in the study for 7 years of follow-up. All participants unstable angina. All participants were classified into two groups, were followed every 2 years. In the case of any report of cardiovascular namely, middle and old aged, based on their age at baseline events or hospital admissions given by participants or their close relatives, (whether 35–55 or 455 years, respectively). a group of trained nurses tried to find reliable documents describing the events, such as the registry or medical records and death certificates, and carry out secondary interviews or verbal autopsies, and a panel of Comparison between demographic characteristics and CVD specialists consisting of cardiologists and neurologists made the final Baseline characteristics of the total population are presented in decision on the diagnosis of the CVD events. Cardiovascular events were fi Table 1. Participants who developed CVD were more likely to be de ned as either coronary heart disease, including fatal and non-fatal men (P-value = 0.034), to be 455 years (to be older) myocardial infarction, sudden cardiac death and unstable angina, or stroke. o The definition of all end points has been extensively explained in a (P-value 0.001) and to have higher values of BMI previous report.24 Waist circumference, blood pressure and body mass (P-valueo0.001) than those without CVD. Furthermore, patients index (BMI) measurements were carried out and lifestyle behaviors, with cardiovascular events were more likely to have lower including smoking status (ever smoke/never smoke), physical activity and educational achievements than those without cardiovascular dietary intake, were asked. Data on physical activity, expressed as events. The results indicated that among the participants with metabolic equivalent task minutes per week, were obtained through an o5 years of education, 8.6% had developed CVD; among the oral questionnaire that included questions on recreation, sport and leisure- participants with 6–12 years and 412 years of education, 6.5% time physical activity. We requested participants to think about all the vigorous and moderate activities they had executed in the last 7 days, and 5.8% had CVD, respectively. Meanwhile, retired subjects had considering the number of days per week and the time spent on these the highest frequency (15%) of CVD events. Legume consumption activities.27 Common dietary intakes were evaluated with the use of a was categorized based on its tertiles at the following boundaries validated 48-item, semiquantitative food frequency questionnaire that was (minimum-maximum): 0–1 times per week (the first tertile), 1–3 designed based on Countrywide Integrated Non-communicable Disease times per week (the second tertile), more than 3 times per week Intervention protocol.28 Participants were asked to report their frequency (the third tertile). CVD events were less among participants in the of consumption of each food item during the previous year. Dietary intake third tertile than others (6.96% vs 7.6% and 9.8%, P-value = 0.019). on the food frequency questionnaire, which was considered as con- founder, was classified into 12 food groups, as follows: (i) fruits, (ii) vegetables, (iii) dairy products, (iv) non-hydrogenated vegetable oils, The relationship between legume consumption and time to CVD (v) legumes, (vi) nuts, (vii) white meat, (viii) grains, (ix) hydrogenated based on modeling vegetable oils, (x) red meat, (xi) processed meat and (xii) sweets and pizza. Hazard ratios (95% CI) for CVD events based on age and legume fi ’ We quanti ed participants intake of foods from these groups and divided intake were reported (Table 2). Those participants in the first tertile the participants into quintiles according to their consumption. Individuals fi were considered as reference. Risk estimation differed significantly in the two highest intake quintiles for (i) to (vii) were classi ed as having a fi healthy diet and were given a score of 1 for each food group, whereas based on the age. When participants were strati ed by age at those in the other intake quintiles of the food groups were given a score of entry into two groups (middle and old aged), the inverse 0. For unhealthy food groups, the opposite was done. The total dietary significant association was detected in the old category. Accord- score was calculated as the sum of the scores given for all 12 food groups ingly, in the old group, those in the third tertile had a lower risk of 27 (ranging from 0 to 12 for each individual). CVD (model 1 (unadjusted model): 0.64; 95% CI, 0.44–0.93). Likewise, hazard ratio and 95% CI in model 2 (unadjusted model Statistical analysis along with shared frailty) were 0.67 and 95% CI, 0.46–0.97. Data were analyzed by STATA software (Stata/IC 12.0, Stata Corp LP, Furthermore, when we considered the total population, in model 2, College Station, TX, USA). Quantitative variables were expressed as significant and inverse associations were detected. Indeed, those mean ± s.d. and qualitative variables were expressed as counts (percent). who were in the second tertile and the third tertile of legume Qualitative variables were compared with CVD using the Χ2 and intake had a lower risk of CVD in this model (based on HR (95% CI): quantitative variables by using independent t-test or Mann–Whitney test 0.79 (0.63, 0.993) and 0.73 (0.55, 0.96), respectively). In model 2, (if the normality assumption was not held). Subjects were recruited at variances (SE) of shared frailty are equal to 0.02 (0.05) in those 35– multiple clusters. Factors varying across clusters may exert a powerful 55 years old, 0.11 (0.07) in those 455 years old and 0.04 (0.03) in fl independent in uence on study outcomes. Thus, explanatory variables the total population. Variance parameter accounts for variation cannot explain all the variability in the observed time to event. The shared due to unobservable factors. A nonsignificant variance would frailty model is a practical and appealing tool for addressing heterogeneity between clusters in the analysis of multistage cluster sampling with time to indicate that the frailty component does not contribute to the event outcome.29,30 The present shared frailty groups consisted of 60 areas model. Thus, the Cox proportional hazard model was used for based on individuals from the same cluster of urban or rural areas of the analysis. Significant shared frailty shows the unobservable three selected districts. If the frailty effect was not statistically significant, it common factors within clusters of subjects. Results of the was excluded and a conventional Cox proportional hazard model was used likelihood ratio test for checking the equality of variance to fit to the data. In shared frailty, the tertile of legumes was considered as parameter with zero are reported in Table 2. The results showed

European Journal of Clinical Nutrition (2016) 217 – 221 © 2016 Macmillan Publishers Limited Association of legume intake with cardiovascular disease risk F Nouri et al 219 DISCUSSION Table 1. Baseline demographic, socioeconomic, lifestyle and anthropometric characteristics of the study population Adding shared frailty to survival model is a way to account for variation, due to unobservable factors within clusters. The present Characteristics Cardiovascular disease P-value study showed a strong relationship between dietary intake of (CVD) legumes and the risk of CVD events in the Iranian population older than 55 years, using the shared frailty survival model. In fact, the No Yes hazard ratios for CVD events decreased with the highest intake of legumes compared with the lowest consumers. Also, consuming Age at baseline 50.07 ± 11.45 57.74 ± 11.47 o0.001 legumes three times or more per week compared with less than a (years) once a week after controlling for all considered confounders such b – as sex, income, marital status, education, job, ever smoke, leisure- Age category 35 55 years 3472 (95.07) 180 (4.93) o0.001 time physical activity, legume-adjusted dietary score, anti- 4 55 years 1499 (85.85) 247 (14.15) hyperlipidemia medicine and BMI in the full adjusted model was b associated with 34% lower risk of CVD in the old-age population. Sex Male 2400 (91.29) 229 (8.71) 0.034 Female 2571 (92.85) 198 (7.15) This conventional full adjusted model does not include unob- served factors leading to an extra layer of heterogeneity and b Marital status Single 419 (88.40) 55 (11.60) 0.002 greater variability in survival times. However, there was no Married 4552 (92.45) 372 (7.55) signiﬁcant association between the frequency of consuming

b legumes and CVD events in the middle-aged population. More- Occupation Housewife 2455 (92.50) 199 (7.50) over, inverse association was observed between the frequencies Working 2143 (92.93) 163 (7.07) o0.001 Retired 367 (84.95) 65 (15.05) of legume intake and CVD events in some models in the total population. Similar results were observed in many studies. For Educationb 0–5 years 3510 (91.43) 329 (8.57) example, Bazzano et al. observed signiﬁcant inverse relationship 6–12 years 1145 (93.55) 79 (6.45) 0.02 between legume intake and risk of coronary heart disease and 412 years 308 (94.19) 19 (5.81) CVD after adjustment for establishing CVD risk factors in 9632

b subjects free of CVD at baseline. They showed that legume Income Low 4165 (92.19) 353 (7.81) consumption four times or more per week compared with less Medium 617 (92.37) 51 (7.63) 0.51 than once a week was associated with 22% lower risk of coronary High 43 (87.76) 6 (12.24) heart disease and 11% lower risk of CVD.15 In another study, in five b cohorts in Japan, Sweden, Greece and Australia (Anglo-Celtic, Ever smoke No 3895 (92.58) 312 (7.42) 0.013 Yes 1070 (90.37) 114 (9.63) Greeks) among long-lived older people (older than 70 years), higher legume intake was the most protective dietary predictor of Dietary scorea 5.06 ± 1.74 5.27 ± 1.75 0.016 survival among the older, regardless of their ethnicity. The significance of legumes persisted even after controlling for age Leisure-time 74.84 ± 85.28 67.95 ± 84.48 0.11 31 a at enrollment (in 5 years intervals), gender and smoking. physical activity Increased low-density lipoprotein cholesterol levels; high fasting BMIa 26.61 ± 4.42 27.59 ± 4.84 o0.001 serum triglyceride; low high-density lipoprotein cholesterol values; hypertension; diabetes; obesity (increased BMI); and aging are the Tertile of legumeb,c Tertile 1 1056 (90.18) 115 (9.82) major risk factors for CVD events that may be favorably altered by Tertile 2 2698 (92.43) 221 (7.57) 0.019 dietary intervention. It is generally accepted that legumes and Tertile 3 1217 (93.04) 91 (6.96) whole grains have the potential to favorably modify these risk a ± factors and may protect against CVD through various Data are expressed as mean s.d. and P-value obtained from t-test or 13–17 Mann–Whitney test (if required). bData are expressed as number (%) and mechanisms. A meta-analysis of 11 clinical intervention trials P-value obtained from Χ2 test. cTertile 1: 0–1 times per week, tertile 2: 1–3 involving legumes (other than soybeans) found overall a 6.2% times per week, tertile 3: 43 times per week. lowering of low-density lipoprotein cholesterol and 22% lowering of triglycerides.13 Moreover, one investigation revealed that in addition to the blood lipid lowering, legumes and soy have bioactive components that improve vascular function. Further- that in the older group and the total population the frailty effect more, increased consumption of legumes and soy lowers CVD risk was statistically significant and model 2 was a conventional model simply by reducing the intake of protein sources high in saturated 32 33 contrary to model 1. The hazard ratio (95% CI) in model 2 was 0.67 fatty acids. Jang et al. randomized 76 men with heart disease (0.46–0.97) in the old-aged group and 0.73 (0.55, 0.96) in the total to either refined rice or legume powder for 16 weeks and showed population. These are the estimated hazard ratios for two that the whole-grain and legume diet reduced plasma insulin and individuals from the same cluster of urban or rural areas of the glucose by 14%. An increasing number of observational findings three districts, Isfahan, Arak and Najaf Abad, having the same have reported that a higher intake of dietary fiber existing in the frailty in which one has an intake of legume in the third tertile and legumes, whole grains, fruits and vegetables is associated with a the other has an intake of legume in the first tertile. Because the lower risk of CVD and myocardial infarction. Mozaffarian's previous frailty effect was not statistically significant in model 4, it was report on cereal, fruit and vegetable fiber intake and the risk of excluded and the Cox proportional hazard model was used to CVD in older individuals (65 years and older) proposed that cereal analyze the data (model 3). After further adjustment for BMI, a fiber consumption was inversely associated with risk of CVD, with reduction was shown in the risk of CVD. Because of the 21% lower risk after adjustment for age, sex, education, diabetes, nonsignificant shared frailty variance in model 6, hazard ratio smoking, physical activity, fruit and vegetable fiber consumption. (95% CI) in the conventional multiple model adjusted by It has been shown that soluble fiber affects serum lipid levels, confounders, were 0.66 (0.45–0.98) in the third tertile compared blood pressure and insulin resistance.34 Moreover, dietary fiber with the first tertile of legume consumption in an old-aged group may affect fibrinolysis and coagulation that may be important in (model 5). The risk of CVD was significantly negatively associated the setting of established atherosclerotic plaques.35–37 In addition, with frequency of legume consumption in this situation. legumes are important dietary sources of antioxidant vitamins and

Table 2. Hazard ratios and 95% CI of tertile of legume consumption for cardiovascular disease (CVD)

Cardiovascular disease (CVD)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6

Tertiles of legume Age category ⩽ 55 years Tertile 1 Ref. Ref. Ref. Ref. Ref. Ref. Tertile 2 0.91 (0.62, 1.33) 0.91 (0.62, 1.33) 0.84 (0.57, 1.24) 0.85 (0.58, 1.25) 0.87 (0.59, 1.28) 0.87 (0.59, 1.28) Tertile 3 0.94 (0.61, 1.45) 0.96 (0.62, 1.48) 0.88 (0.57, 1.36) 0.91 (0.58, 1.41) 0.92 (0.59, 1.43) 0.92 (0.59, 1.43)

Tertiles of legume Age category455 years Tertile 1 Ref. Ref. Ref. Ref. Ref. Ref. Tertile 2 0.82 (0.62, 1.09) 0.86 (0.65, 1.15) 0.84 (0.62, 1.14) 0.87 (0.64, 1.17) 0.88 (0.65, 1.19) 0.89 (0.66, 1.21) Tertile 3 0.64 (0.44, 0.93) 0.67 (0.46, 0.97) 0.63 (0.42, 0.93) 0.64 (0.43, 0.95) 0.66 (0.45, 0.98) 0.67 (0.45, 0.99)

Tertiles of legume Total population Tertile 1 Ref. Ref. Ref. Ref. Ref. Ref. Tertile 2 0.77 (0.62, 0.97) 0.79 (0.63, 0.99) 0.86 (0.68, 1.09) 0.88 (0.69, 1.12) 0.89 (0.70, 1.12) 0.89 (0.71, 1.14) Tertile 3 0.69 (0.53, 0.92) 0.73 (0.55, 0.97) 0.75 (0.56, 1.00) 0.78 (0.59, 1.05) 0.79 (0.59, 1.05) 0.81 (0.60, 1.08) Data are expressed by hazard ratio (95% CI). Model 1: unadjusted model. Model 2: unadjusted model along with shared frailty (variance of shared frailty ± s.e. for age groupo55 years: 0.02 ± 0.05, P-value = 0.37, for age group455 years: 0.11 ± 0.07, P-value = 0.014, for overall participant: 0.04 ± 0.03, P-value = 0.048). Model 3: adjusted by age, sex, income, marital status, education, job, ever smoke, leisure-time physical activity, legume adjusted dietary score and anti- hyperlipidemia medicine. Model 4: adjusted by age, sex, income, marital status, education, job, ever smoke, leisure-time physical activity, legume-adjusted dietary score and anti-hyperlipidemia medicine, along with shared frailty (variance of shared frailty ± s.e. for age group o55 years: 0.03 ± 0.06, P-value = 0.26, for age group455 years: 0.06 ± 0.06, P-value = 0.12, for total (all of the participants): 0.04 ± 0.03, P-value = 0.09). Model 5: adjusted by age, sex, income, marital status, education, job, ever smoke, leisure-time physical activity, legume-adjusted dietary score, anti-hyperlipidemia medicine and BMI. Model 6: adjusted by age, sex, income, marital status, education, job, ever smoke, leisure-time physical activity, legume-adjusted dietary score, anti-hyperlipidemia medicine and BMI, along with shared frailty (variance of shared frailty ± s.e. for age group o55 years: 0.003 ± 0.05, P-value = 0.47, for age group 455 years: 0.05 ± 0.06, P-value = 0.18, for total (all of the participants): 0.02 ± 0.03, P-value = 0.20).

unsaturated fatty acids. The consumption of polyunsaturated fatty association between legume consumption and CVD. Further acids and vitamin E is associated with protection from investigations are required to confirm our findings considering coronary artery disease.22 Hosseinpour–Niazi et al. surveyed the longer follow-up duration and total energy intake adjustment association between dietary legume intake and metabolic while considering other comorbidities. syndrome and its components in a case − control study of 80 subjects aged 25–55 years old. The findings showed that, after statistical control for confounders, decreases in triglyceride CONFLICT OF INTEREST concentrations, fasting blood glucose, systolic blood pressure The authors declare no conflict of interest. and increase in high-density lipoprotein cholesterol concentration were observed across increasing quartile categories of legume 38 intake. To our knowledge, there has not been any report ACKNOWLEDGEMENTS regarding the influence of legume intake on the risk of cardiovascular events with the shared frailty model. We would like to acknowledge the ICS staff and collaborators for their contribution. The strengths of this study include the use of a wide range of likely confounding variables, including sex, age, marital status, education level, job, ever smoke, leisure-time physical activity, REFERENCES legume-adjusted dietary score, anti-hyperlipidemia medicine and 1 Murray CJ, Lopez AD. Mortality by cause for eight regions of the world: Global BMI, a large and representative sample, a comprehensive Burden of Disease Study. Lancet 1997; 349: 1269–1276. statistical method with focus on the correlation between clusters 2 World Health Organization. Mortality Estimates by Cause, Age, and Sex for the Year of observations and variation due to unobservable or latent 2008. WHO: Geneva, 2011. 3 Tunstall-Pedoe H. Preventing chronic diseases. A vital investment: WHO global factors common within clusters of subjects. However, some report. Int J Epidemiol 2006; 35: 1107. limitations in the interpretation of the results must be considered. 4 Forman D, Bulwer BE. Cardiovascular disease: optimal approaches to risk factor Our questionnaire was qualitative and did not quantify food modification of diet and lifestyle. Curr Treat Options Cardiovasc Med 2006; 8: 39 consumption, while total energy intake was not measured. Also, 47–57. our study has limitations due to the relatively short duration 5 Willett W. Nutritional Epidemiology. Oxford University Press: New York, NY, USA, without repeated assessment of diet40,41 during the follow-up 1990. period and no other comorbidities. 6 Anderson JW, Smith BM, Washnock CS. Cardiovascular and renal benefits of dry bean and soybean intake. Am J Clin Nutr 1999; 70:464S–474S. 7 Liu S, Buring JE, Sesso HD, Rimm EB, Willett WC, Manson JE. A prospective study of CONCLUSION dietary fiber intake and risk of cardiovascular disease among women. JAmColl Cardiol 2002; 39:49–56. Legumes were found to be healthy food choices. We thus 8 Pietinen P, Rimm EB, Korhonen P, Hartman AM, Willett WC, Albanes D et al. Intake recommend frequent legume consumption in our country. In spite of dietary fiber and risk of coronary heart disease in a cohort of Finnish men. The of a wide level of adjustment in the present study, it is still Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study. Circulation 1996; 94: possible that unmeasured confounders interfere with the 2720–2727.

European Journal of Clinical Nutrition (2016) 217 – 221 © 2016 Macmillan Publishers Limited Association of legume intake with cardiovascular disease risk F Nouri et al 221 9 Moat SJ, Lang D, McDowell IF, Clarke ZL, Madhavan AK, Lewis MJ. Folate, community-based programme for cardiovascular disease prevention and control. homocysteine, endothelial function and cardiovascular disease. J Nutr Biochem Design, methods and initial experience. Acta Cardiol 2003; 58: 309–320. 2004; 15:64–79. 26 Sarrafzadegan N, Baghaei A, Sadri G, Kelishadi R, Malekafzali H, Boshtam M et al. 10 Gums JG. Magnesium in cardiovascular and other disorders. Am J Health Syst Isfahan Healthy Heart Program: Evaluation of comprehensive, community-based Pharm 2004; 61: 1569–1576. interventions for non-communicable disease prevention. Prev Control 2006; 2: 11 Fox CH, Mahoney MC, Ramsoomair D, Carter CA. Magnesium deficiency in African- 73–84. Americans: does it contribute to increased cardiovascular risk factors? J Natl Med 27 Sarrafzadegan N, Kelishadi R, Esmaillzadeh A, Mohammadifard N, Rabiei K, Roo- Assoc 2003; 95:257–262. hafza H et al. Do lifestyle interventions work in developing countries? Findings 12 Saari JT. Copper deficiency and cardiovascular disease: role of peroxidation, from the Isfahan Healthy Heart Program in the Islamic Republic of Iran. Bull World glycation, and nitration. Can J Physiol Pharmacol 2000; 78: 848–855. Health Organ 2009; 87:39–50. 13 Anderson JW, Major AW. Pulses and lipaemia, short- and long-term effect: potential 28 Mohammadifard N, Sarrafzadegan N, Nouri F, Sajjadi F, Alikhasi H, Maghroun M 88 – in the prevention of cardiovascular disease. Br J Nutr 2002; :S263 S271. et al. Using factor analysis to identify dietary patterns in Iranian adults: Isfahan fi 14 Bazzano LA, He J, Ogden LG, Loria CM, Whelton PK. Dietary ber intake and Healthy Heart Program. Int J Public Health 2012; 57: 235–241. reduced risk of coronary heart disease in US men and women: the National Health 29 Clayton DG. A model for association in bivariate life tables and its application in and Nutrition Examination Survey I Epidemiologic Follow-up Study. Arch Intern epidemiological studies of familial tendency in chronic disease incidence. Bio- 163 – Med 2003; : 1897 1904. metrika 1978; 65:141–151. 15 Bazzano LA, He J, Ogden LJ, Loria C, Vupputuri S, Myers L et al. Legume con- 30 Hougaard P. Analysis of Multivariate Survival Data. Springer: New York, 2000. sumption and risk of coronary heart disease in US men and women: NHANES I 31 Darmadi-Blackberry I, Wahlqvist ML, Kouris-Blazos A, Steen B, Lukito W, Horie W Epidemiologic Follow-up Study. Arch Intern Med 2001; 161: 2573–2578. et al. Legumes: the most important dietary predictor of survival in older people of 16 Geil PB, Anderson JW. Nutrition and health implications of dry beans: a review. different ethnicities. Asia Pacific J Clin Nutr 2004; 13:217–220. J Am Coll Nutr 1994; 13: 549–558. 32 Flock MR, Kris-Etherton PM. Dietary guidelines for Americans 2010: implications 17 Serrano J, Goni I. Effects of black bean Phaseolus vulgaris consumption on the for cardiovascular disease. Curr Atheroscler Rep 2011; 13:499–507. nutritional status of Guatemalan population. Arch Latinoam Nutr 2004; 54:36–44. 33 Jang Y, Lee JH, Kim OY, Park HY, Lee SY. Consumption of whole grain and legume 18 Liu S, Stampfer MJ, Hu FB, Giovannucci E, Rimm E, Manson JE et al. Whole-grain powder reduces insulin demand, lipid peroxidation, and plasma homocysteine consumption and risk of coronary heart disease: results from the Nurses’ concentrations in patients with coronary artery disease: randomized controlled Health Study. Am J Clin Nutr 1999; 70: 412–419. clinical trial. Arterioscler Thromb Vasc Biol 2001; 21:2065–2071. 19 Jenkins DJ, Kendall CW, Augustin LS, Mitchell S, Sahye-Pudaruth S, Blanco Mejia S 34 Mozaffarian D, kumanyika SK, Lemaitre RN, Olson JL, Gl Burke, Siscovick DS. et al. Effect of legumes as part of a low glycemic index diet on glycemic control Cereal, fruit and vegetable fiber intake and the risk of cardiovascular disease in and cardiovascular risk factors in type 2 diabetes mellitus: a randomized 289 – controlled trial. Arch Intern Med 2012; 172: 1653–1660. older individuals. JAMA 2003; : 1659 1666. fi 20 Bhattacharya S, Malleshi NG. Physical, chemical and nutritional characteristics of 35 Pereira MA, Pins JJ. Dietary ber and cardiovascular disease: experimental and 2 – premature-processed and matured green legumes. J Food Sci Technol 2012; 49: epidemiologic advances. Curr Atheroscler Rep 2000; :494 502. fi 459–466. 36 Jenkins DJ, Axelsen M, Kendall CW, Augustin LS, Vuksan V, Smith U. Dietary bre 83 21 Han H, Baik BK. Antioxidant activity and phenolic content oflentils (Lens culinaris), lente carbohydrates and the insulin-resistant diseases. Br J Nutr 2000; : – chickpeas (Cicer arietinum L.), peas (Pisum sativum L.) and soybeans (Glycine max), S157 163. and their quantitative changes during processing. Int J Food Sci Technol 2008; 43: 37 Erkkila AT, Lichtenstein AH. Fiber and cardiovascular disease risk: how strong is 21 – 1971–1978. the evidence? J Cardiovasc Nurs 2006; :3 8. 22 Anderson JW, Hanna TJ. Impact of nondigestible carbohydrates on serum lipo- 38 Hosseinpour-Niazi S, Mirmiran P, Amiri Z, Azizi F. Dietary legumes intake and proteins and risk for cardiovascular disease. J Nutr 1999; 129: S1457–S1466. metabolic syndrome and its component in adults. J Endocrinol Metab 2011; 12: 23 Tourlouki E, Matalas AL, Panagiotakos DB. Dietary habits and cardiovascular dis- 594–602. ease risk in middle-aged and elderly populations: a review of evidence. Clin Interv 39 Willett WC, Howe GR, Kushi LH.. Adjustment for total energy intake in epide- Aging 2009; 4: 319–330. miologic studies. Am J Clin Nutr 1997; 65: 1220S–1228S. 24 Sarrafzadegan N, Talaei M, Sadeghi M, Kelishadi R, Oveisgharan S, Mohammadi- 40 Oh S, Hong M. Within-and between-person variation of nutrient intakes of older fard N et al. The Isfahan cohort study: rationale, methods and main findings. people in Korea. Eur J Clin Nutr 1999; 53:625–629. J Hum Hypertens 2011; 25: 545–553. 41 Ogawa K, Tsubono Y, Nishino Y, Watanabe Y, Ohkubo T, Watanabe T et al. Inter- 25 Sarraf-Zadegan N, Sadri G, Malek Afzali H, Baghaei M, Mohammadi Fard N, and intra-individual variation of food and nutrient consumption in a rural Japa- Shahrokhi S et al. Isfahan Healthy Heart Programme: a comprehensive integrated nese population. Eur J Clin Nutr 1999; 53: 781–785.